JP4833983B2 - Voltage source converter - Google Patents

Description

  The present invention relates to a voltage source converter (VSC) having a plurality of self-commutating semiconducting elements. More precisely, the present invention relates to a method and apparatus for controlling a VSC by a modulation signal, such as a pulse width modulation (PWM) signal. In particular, the present invention relates to a VSC having a plurality of semiconductor elements connected in series, and more particularly to a converter station of a high voltage direct current (HVDC) transmission line having such a VSC.

  In this description, the PWM control converter includes an inverter as well as a rectifier. Such converters are used in low power applications, such as motor drive systems, and for example, high voltage direct current (HVDC) transmission systems and static var compensation (STACOM) systems. It can be used in high power applications.

  The converter, and more precisely the voltage source converter, provides an electrical coupling between the DC voltage system and the AC voltage system having one or more phases. Depending on the direction of power, the converter can either function as a rectifier (function to supply power from the AC system to the DC system) or an inverter (function to supply power from the DC system to the AC system). Have. For example, the converter can be used for variable speed control of synchronous or asynchronous rotating machines as well as transmission of high voltage direct current (HVDC) over long distances.

  The simplest transducer has a two-level bridge consisting of two valves. Each valve has one or more switches. The three-phase converter thus has a bridge with six valves, where each valve has at least one switch. The switch has a turn-off device and a diode connected antiparallel thereto. With this arrangement, the current is stopped in a controllable manner in one direction, but is free to pass in the opposite direction. For high voltage applications, each valve has a plurality of series switches with such turn-off devices and anti-parallel diodes.

  Since the load is an inductive type, a diode (called a “free wheeling diode”) is connected in parallel with the switch to allow the load current to flow when the corresponding switch is opened. Need to be deployed. A further development of the two-level converter is a three-level converter, which requires six special diodes. This transducer is also referred to as a neutral point clamped (NPC) transducer bridge.

  As an example, when one bridge of a two-level converter is used, the converter's AC output voltage, strength, phase angle, fundamental frequency, and harmonic distortion are bridges connected to the same phase. The top two valves are controlled by alternately switching on and switching off. Thereby, the alternating current is controlled in a desired state. A pulse signal for controlling the switch is generated based on a selected pulse width modulation (PWM) method.

  The PWM method is rich in variety. The most frequently used methods are, for example, carrier-based PWM, such as sinusoidal pulse width modulation (SPWM), and, for example, optimal pulse width modulation (OPWM). This is carrierless PWM. The modulation techniques in the prior art are based on the assumption that the switching elements of the converter operate in an ideal state, i.e. they are switched on or off exactly at the moment required by the control. These are referred to as ideal switching moments in the text below. In reality, however, the output voltage waveform of the converter deviates from what the control originally required.

  The first reason is that switching devices are not ideal. Switching devices have a lag in response to their control signals during turn-on and turn-off switching, respectively. This reaction delay depends on the type of semiconductor, its current and voltage rating, the control waveform at the gate electrode, the temperature of the device, and in particular the actual current that is switched on and off.

  The second reason is the blank time (ie, “dead time”), which is the time between the first valve opening (turn-off) command and the second valve closing (turn-on) command of the same bridge. In between, it must be inserted. Due to the presence of the blank time, the two valves of the transducer bridge are not closed at the same time, thus preventing a short circuit.

  The third reason is that it contributes to the deformation of the output voltage, and there is a difference in the rate of voltage increase and decrease (dv / dt) across the switching device between turn-off and turn-on. . This may be due to the presence of a spike voltage absorption circuit or parasitic capacitance in the diode. This deformation is particularly noticeable when the switching current is low.

  Based on the reasons listed above, there is a delay between the switching command and the actual switching event. In order to realize the actual switching event corresponding to the ideal switching moment, the switching command must be sent in advance. Thus, for every switching, the valve action time must be taken into account. Valve action time is defined in the text below as the time difference between an actual switching command and its actual switching event.

  Thus, the action time has fluctuations due to switching device response delay, blank time, and low voltage increase and decrease rates (dv / dt). As a result of these parameter variations, non-linear errors occur between the commanded voltage and the actual converter output voltage. This not only results in extra low order harmonics (e.g., 5th and 7th harmonics), but sometimes also introduces control system instability problems. Therefore, many attempts have been made so far to correct or compensate for these errors.

  From US Pat. No. 5,991,176, a method and a device for processing a PWM wave are already known. The purpose of this method is to reduce or eliminate the effect of blank time (called dead time) in the inverter or control rectifier. A known inverter is controlled by a modulator and discriminator. The role of the modulator is to create a set wave, whereas the role of the discriminator is to divide this wave into multiple waves that are each intended to control various switches. It is possible to do. The purpose of the discriminator is to delay the closing of the corresponding switch, so that when the command to close one switch is given, it is always ensured that the opposite switch is already open. Will come to be.

  The known method suggests using two modified control set signals: one for the case where the current is the output current and the other one. , For the case where the current is the input current. It is the direction of the current in the load that determines which of the two modified setting signals is used. In this way, the switching command compensates for the blank time.

  From US Pat. No. 6,535,402, a method for dead time adaptive compensation for inverters and converters is already known. The purpose of this method is to avoid current distortion and torque ripple of a motor driven by such an inverter by compensating for the effects of dead time. This document recognizes the difficulty of measuring the zero crossing of the current and proposes adding a bias current to the distorted current. Thereby, the time when the current has passed the bias level of the current is determined. The second dead time compensation is derived from the bias level current crossing and is added to the first dead time compensation of the PWM signal.

  A known method for correcting the error between the commanded voltage and the actual output voltage in the prior art is based on current measurements. Thus, known methods are based on current switching measurements. A feed forward type compensation has been proposed that only corrects the average voltage error due to blank time or low dv / dt. Errors due to the response time of the switching device are not taken into account. There is no confirmation as to whether feedback control or switching device turn-on or turn-off occurred at the exact moment required by the control. In addition, the method described in US-6,535,402 requires additional hardware components, which can be very expensive in high power applications.

  The methods known from the prior art work well enough under certain operating conditions. In other operating conditions, they cannot function properly. One such case is when the switching frequency is low and the inductance is low, in which case very high current ripple may occur. Typically, in STATCOM and HVDC high power applications, the converter is connected directly to the grid. In such a situation, high switching current ripple will occur. In such a case, it is clear that the current direction is different from one switching instant to the next.

  It would also be possible to use the current predicted at the next switching instant in order to pre-estimate the action time for the next switching. However, it is very difficult to ensure the accuracy of the predicted current. This is because the accuracy of the predicted current depends not only on the reference voltage of the transducer, the accuracy of the measured current and the measured voltage, but also on the calculation speed of the control process.

In high power applications such as HVDC and STATCOM, low order harmonics cause very high costs for the filtering device. Thus, a new control method is required. The method is to enable the realization of high precision switching for voltage source converters used in high power applications, thereby eliminating the effects of the errors mentioned above.
US Pat. No. 5,991,176 US Pat. No. 6,535,402

It is a primary object of the present invention to provide a method and apparatus for controlling a voltage source converter, which increases the accuracy of switching control and provides the error discussed above. The effect of.
A second object of the present invention is to provide a method and apparatus that can eliminate the problems of low order harmonics (eg, 5th and 7th harmonics) and system control instability.

  A further objective is to determine the valve action time with high accuracy. A still further object is to reduce converters with high current ripple in power systems, such as in high power applications, and low current ripple, as in drive systems and other applications. It is to provide an appropriate method for both of the converters having the same. A further object is to provide a method that does not require additional hardware and does not depend on whether the information used is from current measurement or voltage measurement.

  These objects are achieved by an apparatus characterized by the features of independent claim 1, by a method characterized by the steps of independent claim 7, or by a computer program characterized by the features of independent claim 10. The Preferred embodiments are set forth in the dependent claims.

  According to the present invention, the actual switching event is detected and the action time is adjusted by comparing the ideal switching instant with the detected actual switching event. The ideal switching instant time is subtracted from the actual switching event time and added to the current action time to form an adjusted action time.

  In this way, if the calculated difference is positive, the action time is increased, and if the difference is negative, the action time is decreased. If there is no difference between the ideal switching instant and the actual switching event, no adjustment of the action time is necessary.

  The time difference between the ideal switching instant and the actual switching event from the first pulse can be used to correct the action time for the next pulse. Doing so creates two main considerations. First, the hardware performance for calculating the difference and the adjustment required between two adjacent pulses is enormous. Second, the operating condition for switching the first pulse may not be the same as the operating condition for switching the second pulse. In this way, the action times can be different and the adjustment can be worse than if it only calculates the moment of sending the switching command.

  According to the present invention, the adjusted action time for the selected pulse in the first period of the fundamental frequency corrects the actual switching command for the same pulse in the subsequent period of the fundamental frequency. Used for. In this way, the information obtained from the first period is used to determine the switching command in the subsequent period. By storing the action time for the pulses in the first period of the fundamental frequency, a long time is obtained to calculate the adjustment of the switching command for the next period of the fundamental frequency.

  In this way, hardware performance requirements are reduced. By adjusting the action time of the same pulse in adjacent periods, the component's response delay and variations related to its operating conditions are taken into account. The reason is that the switching operating conditions will be the same for the corresponding pulses in adjacent periods of the fundamental frequency.

According to a first aspect of the present invention, the above object is achieved by a method for controlling a VSC with a PWM pulse signal having an ideal switching instant for each switching pulse.
In this method, the actual switching event for the selected switching pulse in the first period of the fundamental frequency is detected; by comparing the actual switching event with the ideal switching moment, Adjust the action time for the selected switching pulse; modify the switching command of the corresponding pulse in the subsequent period of the fundamental frequency with this adjusted action time.

  For each corresponding pulse in adjacent periods of the fundamental frequency, the operating conditions are primarily the same. The current load is the same and the position in the cycle is also the same. In this way, the reaction times for two corresponding pulses in different periods will be the same. Such an adaptive method self-adjusts the uncertainty in determining the reaction time of the semiconductor depending on the operating conditions of the current. This method can be applied not only to a stationary system but also to a variable frequency system, and is particularly effective when the fluctuation is slow.

  In a preferred embodiment of the present invention, the average action time is calculated for each pulse in the period from the action time of the equivalent pulse in the preceding period. In this way, this stored value is the average of the previous value and the new value. This calculation method may be either a linear average value or an exponential average value.

  In a further preferred embodiment of the invention, the determination of the switching event is evaluated by measuring the voltage across the electrodes of the semiconductor element.

  By adjusting the actual command instant for each pulse in the period of the fundamental frequency from the information of the same pulse in the previous period, the voltage change of the valve occurs exactly at the instant requested by the control. It will be.

  An advantage of the present invention is that low order harmonics are kept to a minimum level. This will greatly reduce the cost of the filter. Another advantage is that control instability for converters with OPWM is eliminated.

In a second aspect of the present invention, the above object is achieved by a control device that provides a pulse width modulation (PWM) signal for controlling the valve of the transducer bridge. The controller comprises sensing means for sensing actual switching events of the semiconductor device; and memory means for calculating and storing an action time for each pulse in the period of the fundamental frequency. And computer means for modifying the actual switching command of the corresponding pulse in the subsequent period of the fundamental frequency.
The apparatus further comprises signal generating means for producing information and transferring information to and from the computer means; sensing means; and semiconductor elements in the transducer.

  In a preferred embodiment of the present invention, the PWM is a carrierless PWM, eg, optimal pulse width modulation (OPWM), or a carrier-based PWM, eg, a sinusoidal pulse. It is width modulation (SPWM: sinusoidal pulse width modulation).

  In the third aspect of the present invention, these objectives provide a method for correcting the actual command instant of a pulse in a fundamental frequency period with information from an equivalent pulse in a preceding period of the fundamental frequency. Implemented by a computer program product having instructions to the device for execution. The computer program also calculates the action time for each switching pulse.

  Other features and advantages of the present invention will become more apparent to those skilled in the art from the following detailed description using the accompanying drawings.

  A two-level converter bridge is shown by way of example in FIG. FIG. 1a shows the entire three-phase forced commutation bridge, and FIG. 1b shows one phase part of the bridge. The bridge portion has a first valve V1 and a second valve V2, and has a lower DC terminal Udn and an upper DC terminal UdP. Each valve has at least one switching device including a self-excited semiconductor element and a diode element connected anti-parallel thereto. In the embodiment shown here, the self-excited semiconductor element has an IGBT. This bridge has an AC terminal Uac having an AC current i.

  When operating the transducer, the blanking time (or “dead time”) is between the first valve opening (turn-off) command and the second valve closing (turn-on) command, and It must be inserted during the reverse relationship. The reason is that these two valves of the transducer bridge must not be closed at the same time to prevent short circuits. The effect of blank time is shown in FIG. The first waveform 1 is an ideal switching pulse. The second waveform 2 is a command pulse for the first valve V1, and the third waveform 3 is a command pulse for the second valve V2. The fourth waveform 4 is the resulting voltage Uac. The blank time is indicated by tb.

FIG. 2 shows that both the phase position and the voltage time domain (which determines the intensity) are different from the ideal pulse (commanded output voltage).
As shown in the figure, a positive current value is defined as the input current. If the current is positive, the IGBT of the second valve V2 and the diode of the first valve V1 carry the current. In this case, the current entering the second valve V2 and the voltage across it will change almost instantaneously when a turn-off command is received by the gate device. However, when a switching off command is sent to the first valve V1, the current entering the first valve V1 and the voltage across it will not change. Changes in the current and voltage of the first valve V1 occur only when the second valve V2 receives a turn-on command. As a result, the voltage at the AC terminal will differ from the voltage required by the control. This is shown by comparing the waveform of the ideal pulse 1 with the resulting AC terminal voltage 4.

  If the current is negative, the diode in the second valve V2 and the IGBT in the first valve V1 will conduct the current. A voltage error will be created when a switching off command is sent to the second valve V2, and the resulting AC terminal voltage is the fifth waveform 5 in FIG. As shown.

  When the current intensity is low, the direction of the current can be different from one switching event to the next. As a result, the diodes in both the first valve V1 and the second valve V2 may conduct current during switching off. That is, when the first valve V1 is switched off, the current is negative, and when the second valve V2 is switched off, the current changes to positive. In this case, assuming that the switching device has ideal switching behavior, the voltage at the AC terminal will be as shown by the sixth waveform 6 in FIG.

  In addition, there is a possibility that the IGBTs in both the first valve V1 and the second valve V2 conduct current during switching off. Under this condition, the voltage at the AC terminal is as shown by the seventh waveform 7 in FIG. 2, assuming that the switching device has an ideal switching behavior.

  Therefore, when the current intensity is large, it is clear that both the phase position and the voltage time domain are different from the output voltage commanded from the control. If the current intensity is small, the phase position may be different from the command. However, the voltage time domain appears to be the same as requested from the control. However, it should be noted that the voltage increases more slowly at low current turn-off than at high current.

  As an example, FIG. 3 shows different switching off currents and their corresponding voltages across the valve during the switching off process. The derivative of the voltage is clearly lower when the switching current is 100A than when the switching current is 2500A. This low voltage derivative will also contribute to the voltage error relative to the voltage commanded by the control.

  Switching is affected by the non-linear behavior of the semiconductor device depending on the switching current. For two adjacent pulses, these conditions are rarely the same, especially for high power applications where the switching frequency is low. Therefore, the switching time of the semiconductor device is not the same for two adjacent pulses. Action time is affected as such because it includes blank time in addition to switching time.

  This means that the adaptive calculation of action time for subsequent pulses from the previous pulse information does not contribute to increasing the accuracy of the resulting switching event. . In FIG. 4, the current 5 and pulse signal 6 to the upper valve are plotted against time. From this it is clear that the direction of the current is different between one switching event and the next.

  The present invention is applicable to the phase leg of the high voltage converter circuit, which is shown schematically in FIG. There are usually three phase legs, which have a common DC capacitor 13 in a plant connected to a network of three-phase alternating current. This is in a conventional manner a plurality of power semiconductor devices 11 connected in series (here in the form of an IGBT) and so-called free-wheeling diodes connected in antiparallel to each such device. 12. The number of power semiconductor devices connected in series is actually much larger than that shown in FIG.

  The series connection of the power semiconductor devices is connected to a DC capacitor 13, whereas the phase terminal 14 between the power semiconductor devices is connected to the phase of the AC voltage network, for example by a phase reactor 15. In FIG. 5, the power semiconductor device having a diode disposed on the upper side of the phase terminal 14 forms an IGBT valve, and the power semiconductor device disposed on the lower side forms another IGBT valve.

  All power semiconductor devices in the IGBT valve are turned on simultaneously by signals from the drive 16 (each schematically shown) so that when a positive potential is desired at the phase terminal 14, the first When the power semiconductor device in one IGBT valve conducts current and a negative potential is desired at phase terminal 14, the power semiconductor device in the second IGBT valve will conduct current.

  A DC voltage across the DC capacitor 13 can be used to generate a voltage at the phase terminal 14 by controlling the power semiconductor device based on the determined pulse width modulation pattern (PWM). The basic component of the voltage is an alternating voltage with the desired strength, frequency and phase position. Such control is effected by sending control pulses from the control device 17 to the different drive devices, which are usually performed via optical fibers. In FIG. 5, there is a first optical fiber 9 and a second redundant optical fiber 10.

  The exchange of information between the control device 17 and the drive device 16 is bidirectional communication via an optical fiber. A switching command is sent from the control device 17 to the drive device 16. An indication signal of the switching event can be sent back from the driving device 16 to the control device 17. The control device 17 is arranged at a low voltage potential but is electrically isolated from the driving device 16. The driving device is arranged at a high voltage potential. A switching event instruction signal is generated in the drive controller.

  There are several factors that affect the delay from switch command to actual switching. Switching devices are not ideal, and the switching behavior is highly dependent on the characteristics of the gate driver. Switching devices react late to their control signals during turn-on and turn-off. The delay time depends on the type of semiconductor, its current and voltage rating, the control waveform at the gate electrode, the temperature of the device, and in particular the actual current that is switched on and off. FIG. 6 shows the dependence of the switching delay on the current.

  As shown by FIG. 6, the direction of the current is the most important parameter. The reason is that at the moment of switching, if current is flowing in the IGBT or in the diode, a different current direction is determined. As discussed above, a “dead time” or blank time must be inserted between the first valve turn-off command and the second valve turn-on command. The blank time dominates the delay of the switching operation depending on the current.

  In order to cause actual switching events to occur at ideal switching moments, there is a delay in the response of the switching device and fluctuations in the slow rate of increase and decrease (dv / dt) of the voltage. Must be sent in advance. However, inaccuracy problems arise if the actual switching event does not occur exactly at the ideal switching moment.

  The first result of this switching inaccuracy is that additional lower order harmonics (eg, 5th and 7th order harmonics) are generated. The second result is that instability problems can occur in the control of the system. This is because there is a non-linear error between the commanded voltage and the actual converter output voltage. According to the present invention, an actual switching event is detected and the time difference between the actual switching command and the actual switching event is evaluated online, and accordingly the same frequency in the next period of the fundamental frequency. By adjusting the actual switching command of the pulse, this non-linear error is removed. This function is adequately independent of the direction and intensity of the current.

  The first way to detect actual switching events is to use the measured voltage. By using a voltage divider, the strength of the voltage across the electrode of one power semiconductor device in the valve is measured and compared to a predetermined reference value during the switching off process.

  As shown in FIG. 7, the instant at which the measured voltage 32 passes the reference value 33 is considered an actual switching event. At the moment of the switching event, signal 34 is generated in the gate controller of the semiconductor device. This signal is sent back to the valve control to indicate the moment of the actual switching event. If several individual semiconductor devices fail, some of such signals are sent from different semiconductor devices to their corresponding valve controls. Within valve control, the time from sending the switch-off command 31 to receiving an indication of the actual switching event 34 is stored, which corresponds to the corresponding switching-off in the next period of the fundamental frequency It will be used when adjusting the command.

  According to a preferred embodiment, the reference voltage is equal to about half of the steady state voltage during switching off status.

  A second way to determine the actual switching event is to measure the measured current. Alternating current is measured and is already used for system control and protection. The measured current is sent as an input to the valve control. Relationship between switching current and time delay (from switching off command to actual switching off event) for a particular type of semiconductor device with a particular gating device and control (and a given blank time) Can be obtained through switching tests.

  FIG. 6 shows a functional relationship between the switching current and the time delay as an example. The resulting function is installed as either a table or an equivalent non-linear function in the valve control process. For each measured switching current, the corresponding time delay can be evaluated using a table or non-linear function 41 (as shown in FIG. 8). The estimated time delay for each switching off command is stored and will be used in adjusting the corresponding switching off command in the next period of the fundamental frequency.

The general concept of the first embodiment of the control method and apparatus according to the present invention is shown in FIG. In this embodiment, the pulse control processor PCP compensates for delays that occur during valve switching using adaptive controls. The drive with valve control unit VCU detects the resulting switching event of pulse t _n ¹ of pulse train 19 to control the voltage source converter valve to form fundamental frequency 18.

The pulse signal 20 carrying this information is sent to a pulse control processor (PCP) included in the control device. The PCP also receives a control pulse CP representing the executed switching command. The PCP compares the pulse signal 20 with the control pulse CP to calculate the response time for the pulse t _n ¹ , that is, the delay time from the transmission of the switch command to the resulting switching event. The calculated reaction times 21 for all pulses in the fundamental frequency period are stored in the memory M.
The pulse width modulation controller, represented by block OPWM, sends a pulse signal 22 representing the switching command (ie, ideal switching command) requested by system control. A signal 23 representing the calculated reaction time for the pulse t _n ² is added to the command signal 22 by the adding means 24 to form a new command signal 25. This new command signal is intended to cause an actual switching event to occur at the desired moment. The new command signal 25 is sent to the control pulse creator C to generate a switching command for the next switching.

  Usually, the overall control of the converter in an HVDC application is divided into three main parts. First, there is system control that controls active power / DC voltage, reactive power / AC voltage, and AC current. The desired or ideal pulse is generated from the system control. Secondly, there is a valve control, which corresponds to member 17 in FIG. Third, there is a drive control device, which corresponds to the member 16 in FIG.

The general concept of a second embodiment of the control method and apparatus according to the present invention is shown in FIG. In this embodiment, the reaction time for pulse t _n ¹ , represented by signal 26, is evaluated using the measured alternating current and function block 41. This has been described above and is shown in FIG. The calculated reaction times 21 for all pulses in the fundamental frequency period are stored in the memory M.

A signal 23 representing the calculated reaction time for the pulse t _n ² is added to the command signal 22 by the adding means 24 to form a new command signal 25. This new command signal is intended to cause an actual switching event to occur at the desired moment.

  While the above are advantageous embodiments, the present invention should not be limited to the embodiments shown herein as examples. The main idea behind the present invention is to use information from one switching pulse in the first period of the fundamental frequency harmonic period to switch the equivalent pulse in the next period. Is to control. The resulting switching event determination can thus be evaluated from either voltage or current measurements. Other detail changes will also become apparent to those skilled in the art upon review of the guidance provided herein. Such modifications are included within the scope of the present invention.

FIG. 1a is a diagram of a transducer; FIG. 1b is a general diagram of a bridge of two-level transducers. FIG. 2 is a graph showing ideal pulses, corresponding pulses to the upper and lower valves, and the resulting voltage. FIG. 3 is a diagram showing different switching-off behaviors. FIG. 4 is a diagram showing current ripple. FIG. 5 is a diagram illustrating a phase leg of a high voltage converter circuit. FIG. 6 shows the switching event delay as a function of current. FIG. 7 is a diagram illustrating voltage detection of a switching event. FIG. 8 is a diagram showing current detection of switching event delay. FIG. 9 is a block diagram of a first embodiment of a control method and apparatus according to the present invention. FIG. 10 is a block diagram of a second embodiment of the control method and apparatus according to the present invention.

Claims (13)

An apparatus for controlling a voltage source converter having at least two bridges (V1, V2) of a self-excited semiconductor element (11) connected in antiparallel to a diode (12),
Means (OPWM) for creating a switching control pulse train (19) to form the fundamental frequency (18);
Means (C) for generating a switching command (CP);
Means for detecting a switching event (VCU);
In a device having
Computer means (PCP) for calculating a reaction time (t1, t2) between a switching command (CP) and a switching event (20) for a selected pulse of the pulse train;
Means (24) for properly compensating for switching commands of equivalent pulses in the period following the harmonic period of the fundamental frequency;
A device characterized by comprising: The apparatus of claim 1 having the following characteristics:
Said means for appropriately compensating the switch command comprises memory means (M) for storing the calculated reaction time of each pulse of the harmonic frequency of the fundamental frequency. Device according to claim 1 or 2 having the following characteristics:
Said means (VCU) for detecting a switching event comprises means for measuring a voltage across the electrode of at least one semiconductor device in the valve. Apparatus according to any one of claims 1 to 3 having the following characteristics:
Said means (OPWM) for creating a switching control pulse train comprises an optimum pulse width modulator. The device according to claim 1, having the following characteristics:
The computer means (PCP) for calculating the reaction time comprises means for calculating an average value for the reaction time of each pulse in the harmonic period of the fundamental frequency. The device of claim 5 having the following characteristics:
The average value has an exponential average value. For controlling a voltage source converter (VSC) having at least two bridges (V1, V2) each including a self-excited semiconductor element (11) connected in antiparallel to a diode (12) and a controller (17) In the method
Arranging a pulse train to form a fundamental frequency;
Defining the moment of sending the switching command of the pulse of the pulse train of the fundamental frequency;
Send a switching command;
Determine the actual switching event;
Compare the actual switching event with the desired switching event;
Adjusting the moment of sending a switching command of an equivalent pulse in the next harmonic period of the fundamental frequency;
A method characterized by. A method for controlling a voltage source converter with a pulse width modulated pulse signal having an ideal switching instant for each switching pulse, comprising:
Determining the actual switching event for the selected switching pulse in the first period of the fundamental frequency;
Adjusting the action time for the selected switching pulse by comparing the ideal switching instant with the actual switching event;
Modify the switching command of the corresponding pulse in the subsequent period of the fundamental frequency;
A method characterized by. 9. A method according to claim 7 or 8 having the following characteristics:
The determination of the actual switching event comprises a measurement of the voltage across the electrodes of at least one semiconductor element. The method of claim 7 having the following characteristics:
Said adjustment at the moment of sending a switching command of an equivalent pulse in the next harmonic period has an adjustment of the blank time.   Computer program product having instructions for a processor (PCP) for evaluating the method according to any one of claims 7 to 10.   12. A computer program product according to claim 11, wherein the product is supplied at least in part via a network such as the Internet.   A computer readable medium comprising a computer program product according to claim 11.

Images